History of Inert Gas in Tankers

Download History of Inert Gas in Tankers

Post on 27-Oct-2015




0 download


The Strange History of Tank Inerting


  • The Strange History of Tank Inerting

    Jack Devanney

    Center for Tankship Excellence, USA, djw1@c4tx.org

    1 The Origins of Tankship Inerting

    On the night of February 4th, 1932, the 10,950 dwt Sun Oil tanker Bidwell was cleaning tanks atMarcus Hook refinery on the Delaware River. She had carried crude oil from Texas to the refineryand was now preparing to take gasoline back. At 12:20 AM, she exploded. The first explosion wasfollowed by three more blasts within 25 minutes. The explosions were heard 20 miles away. 18 crewincluding the Master were killed. The Masters wife was pulled from the frigid waters by a Sun Oilemployee who swam to her rescue.

    Tanker explosions were nothing new. They were by far the most frequent kind of non-war tankercasualty. The first known conversion of a wooden sailing ship to carry oil in bulk was the Charles.In 1869, the Charles was fitted with 59 separate tanks. Three years later she caught fire and waslost. Tanker crews knew what to worry about. In a bit of gallows humor, the crew of the firstsuccessful engine-driven, ocean-going bulk oil tanker, the Gluckauf, nicknamed her the Fleigauf (orBlow Up). The Gluckauf did not blow up, but her sister ship the Vorwarts lasted only four yearsbefore she caught fire at Savona, and had to be sunk.

    Table 1 shows a horribly incomplete list of pre-WW II, non-war tanker fires and explosions.

    Table 1: Some pre-WW II, non-War Tanker ExplosionsYear Ship Dead Vol(m3) Synopsis

    1872 charles sail conversion, 59 leaking tanks, lost by fire1888 fergusons steam conversion, blown to pieces in Rouen1890 vorwarts Gluckauf sister, fire at Savona, towed out, sunk1891 lux 16 fire, and sank, 16 killed, 6 saved1892 petrolea struck by lighting, blew up, near Bordeaux, cargo kero1902 nerite fire 1902-03 lightening Bulysses in Canal, total loss1902 bakunin loaded Callao, fire, Some says repairing1907 silverlip 5 exp in no 4 tank, Bay of Biscay, spore to uk petrol, sunk 5 dead1910 manhatten 19100923 sailed from NY to Algiers loaded. Not heard of since.1911 chesapeake exp 40.2N,48.4W, ny to algiers, abandoned1912 spondilus fire 1912-01-131914 kometa 15 tank explosion, 1914-04-26, 15 killed1915 dakotah fire 1915-10-03, sank 2.4S116.4E, Balikpapan/Europe benzine1917 sebastian explosion, loaded, 1917-05-171922 coylet explosion, loaded, sank off Sand key, 1922-02-081926 volga fire, 1926-01-12, destroyed1926 phoenix 2 DH,riveted tank leak, explosion Boston, 2 killed, Ludwig hurt1928 chuky internal explosion, loaded, sank, 1928-02-151932 bidwell 18 tank cleaning, last cargo crude, 18 killed1933 c. s. petrol 3 internal explosion, sank, 1933-07-14, 3 killed1934 la crescenta 29 suddenly disappeared, 29 killed


  • But Sun Oil had had enough. They knew that they could make tankers much safer by inertingthe cargo tanks The technology was already in use in refineries, and could easily be adapted totankers.1 The process consists of diverting a portion of the boiler exhaust gas to a scrubber, whichis little more than a large shower which cools the gas and extracts most of the sulfur, and thenpumping the combusted, low oxygen gas to the tanks. If the system is properly maintained andoperated, the oxygen content in the tank atmosphere will be less than 5%. Normal air is about21% O2. An oxygen content below 10% will not support combustion regardless of the amount ofhydrocarbon vapor in the gas.

    Sun Oil knew the system would not have prevented the first Bidwell explosion. In those days,it was necessary to enter a tank in order to clean it, which would have required purging an inertedtank with fresh air. However, inerting would have made it far more difficult for the fire to spreadto the other tanks, and likely prevented the follow-on explosions.

    In any event, Sun Oil reacted quickly and aggessively. By the end of 1933, all Sun Oil tankerswere fitted with inert gas systems (IGS). This included the Bidwell which survived. The Bidwellwas later torpedoed in 1942 while loaded 30 miles east of Cape Lookout by the U-160. The torpedostruck midships, burning oil was spilled on deck and killed the 2nd Mate; but the fire did not spreadto the undamaged tanks. The crew was able to put the fire out, and make port under the shipsown power. Sun Oil credits inerting for saving many lives on its ships during World War II, butwe dont have much in the way of details.[3].2 The Bidwell ended up being scrapped in 1965 at age45. Tough ship.

    It wasnt long after Sun had implemented inerting before they realized that proper inertingdramatically reduces steel corrosion rates in the cargo tanks. Corrosion is an oxidation process andproceeds much more slowly in an oxygen depleted environment. Sun Oil found that the need forsteel renewals was drasticially reduced and in some cases eliminated. Typically, an uncoated shipin clean product service has a service life of about 12 years, and 20-25 years in dirty service, albeitwith extensive steel renewals from year 15 on. But the Sun ships were different. The AmericanSun which had been in clean service for about four years, passed her 16 year survey without anyreplacement of steel, and her 20 year survey (at which point it had been in clean or partially cleanservice for 8.5 years) with minimal renewals due to external wastage. The Sabine Sun passed her20 year special survey with no renewals in the cargo spaces. This ship had 2.5 years in clean serviceand another 7 years in part clean service.[2, page 239]

    This was not only an important economic saving, but a big safety and enviromental plus aswell. Tank corrosion can easily result in cargo leaking into ballast tanks and/or pump rooms, whereit becomes a major hazard. Devanney found that the single biggest cause of tanker spillage wasexplosions resulting from cargo being where it shouldnt be.[4, page 90]

    1 In the mid-late 1920s, Chevron (then Socal) experimented with tank ship inerting. They installed systems ona total of 19 ships. But after 20 years they claimed there was no safety advantage no reduction in corrosion, anddiscontinued the effort.[1, page 208] This experience is inconsistent with everything else we know about inerting. Wecan only guess that the systems were not properly maintained, starting a downward spiral, which ended up withthe operators deciding inerting was not worth the effort. In the mid-1970s, Chevron became a strong proponent ofinerting.

    2 Boyle of Sun says During World War II, ships equipped with the system repeatedly proved the value of inerting.Torpedoes would blow out a tank and set it afire, but the surrounding tanks remained safe and sound.[2, page241] We do know that, in addition to the Bidwell, the Sun was torpedoed twice and survived, the Pennsylvania Sunsurvived a torpedoing, as did the Atlantic Sun in 1942. However, four Sun Oil tankers were sunk in World War II:the J. N. Pew (1942-02-21), hit by three torpedoes, the Mercury Sun (1942-05-18a, hit by two torpedoes, the Sunoil(1943-04-05), and the Atlantic Sun (1943-02-17). In both the last two cases, the ship was attacked twice. The Sunoilhad straggled from her convoy with engine problems. She was torpedoed at 0745; again at about 1800 that evening,and it still took two more torpedoes to finish her off. The Atlantic Sun was also sailing alone. The first two torpedoesbroke the ship in two, but there was no fire. The stern section floated and was reboarded by the crew who started tosail her stern first to Halifax, when she was hit by a third torpedo, three hours after the first attack.


  • 2 No Reaction for 30 Years

    Amazingly, neither tanker owners nor tanker regulators took any apparent notice of Suns successwith inerting for nearly 30 years.

    There was no attempt to inert tankers during World War II, despite the fact that the US Navywas well aware of the value of inerting. The US aircraft carriers were inerted (with nitrogen). TheJapanese carriers were not. Several authorities claim this was a major factor in the Pacific War.But tankers which could easily have been inerted for less than a $50,000 per ship were not. Insteadgun crews were put on board which did little more than increase the casualty toll. We dont knowhow many tankermen would have been saved by inerting in World War II. We do know that Hockinglists 264 WW2 sinkings of allied tankers in which considerably more than 4,860 lives were lost.[6]3

    After the war, tanker explosions continued. Table 2 is a very incomplete list of post-war through1960 casualties in which inert gas would probably have made a difference The Sinclair Petroloreexplosion off Brazil resulted in the largest ever oil spill at the time by a factor of two. Most likelycause was cargo leaking into the double bottom.

    Table 2: Some 1945 thru 1960 Casualties where inerting probably would have helpedBased on CTX CDB Version 4.5 as of 2010-04-21T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    19480925 esso salta ? 0 lightning, explosion, sank, no details19510420 esso greensboro M 42 7000 fog, dance of death in Gulf of Mexico, collision, fire, 44 killed19510420 esso suez M 2 0 fog, dance of death in Gulf of Mexico, collision, fire, 44 killed19581019 stanvac japan P 19 0 tank cleaning Arabian Sea, explosion, 19 killed, cause?19600325 mobil astral M 4 0 explosion after load Puerto La Cruz, 4 to 23 killed, cause?19600708 esso portsmouth M 0 0 unloading arm failed Milford Haven, spill, fire, explosion19601206 sinclair petrolore M ? 60000 self-unloading OBO exploded, sank off Brazil, no details19601214 world harmony M 0 collision in Bosporus, three ships on fire, 52 killed, no cause19601214 petar zoranic M 20000 collision in Bosporus, three ships on fire, 52 killed, no cause19601214 tarsus M 52 0 collision in Bosporus, three ships on fire, 52 killed, no cause

    One of the most dramatic of these casualties was the Stanvac Japan, Figure 1. She was almostcertainly tank cleaning when she exploded in the Arabian Sea in October, 1958. The explosionripped the deck off almost all the center tanks and tossed the midships house into the sea. At least19 were killed, including everybody on the bridge. Stanvac was a joint venture between Esso andMobil. CTX has no evidence that either reacted to the casualty in any meaningful way.

    About 1960, BP, which was experiencing very rapid tank corrosion in their ships carrying highsulfur Mideast crude, became interested in inerting as a corrosion control method. BP visited Sun,inspected some ships and the system, and developed their own varient which closely followed theSun design. In April, 1961, they retrofitted the British Skill with an inert gas system (IGS).[5]

    Over the next three years, BP conducted an extensive series of corrosion rate tests on five ships,two of which were non-inerted controls. The numbers were somewhat clouded by teething problems,equipment malfunctions, extended periods during with the tanks were exposed to normal air, andmeasurement difficulties. On the British Sovereign, which had the best performing system, theaverage wastage rate was 0.0042 inches/year as opposed to the control ships which had averagewastage rates of 0.0162 and 0.0156 inches per year respectively. During the test period, the av-erage O2 content in the Sovereigns tanks was 5.4% when the tanks were inerted, but 8.2% whenrepair/inspection time were included, implying the tanks were not inerted close to 18% of the time.Overall the Sovereign results agreed roughly with BP laboratory tests which showed 0.014 in/yearwastage in air and 0.005 in/year at 9% our year.4 The relative improvement in the max wastagewas higher than that for the mean, but meaningful quantitative figures were not derived.

    3 Hocking is not complete and often does not give a casualty toll, or tells us only the number of survivors. It isa safe guess that the number of fatalities in these 264 sinkings is 6,000 or more. And that doesnt count the dead innon-sinking explosions.

    4 We shall see that these numbers can be dramatically improved upon by careful operation of the system.


  • Figure 1: Stanvac Japan after tank cleaning explosion. 19 killed.Source: Ships Nostalgia


  • If the IGS had completely stopped all wastage, we would still expect the Sovereign to have awastage rate about 1/5 that of the control ships, due to the percentage of the time her tanks wereexposed to normal air. In other words, while the tanks were inerted at an average of 5.4% O2, thewastage rates were 1/12th that of the control ships.

    The BP program was pushed at a leisurely pace. From 1963 on all new BP crude carriers werefitted with inert gas systems at build. The program did not extend to most existing tankers. On8th August 1966, the 1952 built, 28,598 dwt British Crown was just finishing loading crude atUmm Said, Qatar when she exploded killing 19 and badly injuring 8. The ship was non-inerted,fitted with gauze flame screens. Ex-BP personnel familiar with the ship believe gas accumulatedaft of the forward house which may have been ignited by a spark from a faulty A/C fan. Heres onequote

    The faulty aircon unit was in the officers smokeroom at the after end of the centre castle.The room was sometimes used as the office during loading and unloading as it exitedstraight out at the flying bridge level. While loading tanks 5 and 6 across, gas couldaccumulate in this area. [John Firmin,www.shipsnostalgia.com,posted 2005-08-01]

    3 VLCCs and Inerting

    Aside from Sun and BP, there was little activity on inerting through the 1960s. The US Navy,Esso, and others dabbled with IGS in the late 50s and early 60s.[1] But the results were eitherdeemed inconclusive or failures. Successful introduction of inerting requires strong direction frommanagement, a hard push at the superintendent level, and committment by the crews. If any ofthese are lacking, the whole thing falls apart as soon as the first problem is encountered,

    Tankers continued to explode. Table 3 is an incomplete list of 1961 through 1969 explosions andfires where IGS probably would have helped. At least 218 people were killed and 185,000 m3 of oilspilled in these casualties.

    And tankers exploded in size. At the beginning of the 1960s, the largest tanker afloat was about100,000 tons. In 1966, the first ship over 200,000 tons deadweight was delivered. Since the presshad started calling the 60,000 and 80,000 tonners built in the late 1950s supertankers. No oneknew what to call these new ships. For want of imagination, they became known as VLCC s (VeryLarge Crude Carriers)

    As Table 3 shows, in the space of three weeks in December, 1969, three nearly new VLCCs hadmassive cargo tank explosions. In all three cases the ships were cleaning empty cargo tanks. As wehave seen, tank cleaning casualties were hardly new. There had been many such fires in the past,most recently Seven Skies which killed four crew. But three such explosions in short order, twoof which were Shell VLCCs, at least got Shells attention.

    Cargo tank cleaning is accomplished by machines that look like and work like enormous lawnsprinklers. These gadgets shoot a revolving high pressure jet of sea water around the tank, in theoryblasting the surfaces clean of oil. Two of the tankers involved, the Marpessa and the Mactra wereShell ships. The third was the brand new Kong Haakon VII. The Marpessa, on her maidenballast leg, sank killing two crewmen. The Mactra, Figure 2, and the Kong Haakon VII, Figure3, had a large portion of their main decks blown away but survived.

    Shell instituted a crash research program and came to the conclusion that the high speed jetsof water impinging on the steel surface of the tank were creating static electricity, in somewhatthe same way that rain drops in a thunderstorm do.[8] When enough static electricity builds up, itproduces a spark in space that is full of hydrocarbon vapor.5 The process is tank sized dependentand didnt make itself totally obvious until tanks grew to VLCC size.

    The obvious solution was inerting. Shell started fitting inert gas systems to their large tankers.Some other VLCC owners did as well. But adoption was slow and largely confined to very large,newbuildings. There was still no movement at all on the regulatory front.

    5 The exact mechanism by which the charged mist is ignited is still a subject of controversy.


  • Figure 2: Mactra deck after tank cleaning explosion. Two killed.Source: Auke Visser, supertankers.topcities.com


  • Table 3: Sixties Casualties where inerting probably would have helpedBased on CTX CDB Version 4.4 as of 2010-03-11T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    19610126 esso durham P 0 0 tank cleaning explosion off Gibraltar, hole in way of No 419640306 bunker hill M 5 0 explosion while tank cleaning, broke in 2, sank, 5 killed19640325 san jacinto Y 1 0 explosion in ballast blamed on falling magnesium anode, 1 killed19640703 bonifaz P 25 0 Fabiola hits Bonifaz off Spain fire, sank. Classic dance of death19650327 nora M 0 2800 collision with Otto N Miller, fog, nil info19650327 otto n miller M 0 0 collision with Otto N Miller, fog, nil info19650523 heimvard M 10 32000 hit Muroran jetty at speed, explosion, 18 killed, total loss19650605 luisa P 32 24000 explosion loading Iran, grounded, capsized, nil other inf19650805 kaizo maru P 13 0 explosion loading Saudi, prob massive spill but no info19650923 barbaros P 15 0 explosion discharging gasoline Izmit, nil other info19660225 anne mildred brovi M 0 19000 N. Sea collision with Pentland in fog, cause?, grounded, CTL19660225 pentland M 0 0 N. Sea collision with Pentland in fog, cause?, grounded, CTL19660312 world liberty M 5 0 collision with World Liberty in Red Sea, conflicting data, cause?19660820 british crown P 19 23000 explosion topping off at Umm Said, prob spark from A/C unit19661211 iphigenia M 3 0 tank explosion in ballast, no inert, 3 killed, sank19670103 esso glasgow P 0 600 Non-inerted T2, tank explosion loading Fawley, ship repaired,19670122 jacob verolme P 1 0 Ore/oiler explosion, loaded iron ore, hotwork on deck, sbt Leak?19670322 circe P 38 0 tank exp Med, possibly loose deck equip, broke in two, 38 lost19680228 mandoil ii M 11 50500 collision off Oregon in hvy fog, probable B encounter. Nil info.19680426 assimi iii M 5 20000 ER fire south of Singapore, sank, nil info on initial cause19680506 islas orcadas M 4 13000 ruptured hose loading gasoline La Plata, 3 ships sunk, spill?19680506 fray luis beltran M 0 0 ruptured hose loading gasoline La Plata, 3 ships sunk, spill?19680506 cutral co M 0 0 ruptured hose loading gasoline La Plata, 3 ships sunk, spill?19681020 sitakund P 3 526 tank exp in English Ch. ballast, probably pre-IG tank cleaning19690724 silja P 20 0 Non-inerted Silja hit by Ville de Majunga off Toulon, fire, sank19691006 seven skies P 4 0 non-inerted tank cleaning explosion NE Singapore19691212 marpessa P 2 0 non-inerted VLCC tank cleaning explosion off Dakar, sank19691229 mactra P 2 0 non-inerted VLCC tank cleaning explosion off Mozambique19691230 kong haakon vii P 0 0 non-inerted VLCC tank cleaning explosion off Liberia


  • Figure 3: Kong Haakon deck after tank cleaning explosion. Obvious similarity to Mactra. Source: Auke Visser, supertankers.topcities.com


  • 4 The USCG Finally Acts

    Finally, in 1974, 42 years after the Bidwell, the United States Coast Guard required inerting, butonly on crude tankers over 100,000 tons built after 1974.6 The regulation applied only to shipstrading to the United States. But this was a key impetus. Many front line owners who requiredthe flexibility to trade to the United States started including IGS in their VLCC newbuildingspecifications. Otherwise, tanker owners continued to ignore inerting.

    On the evening of December 17, 1976, the non-inerted 70,630 dwt tanker Sansinena had justfinished discharging at the Union terminal in Los Angeles. As she began ballasting, tank vaporswere pushed out on deck. Winds were light. At 19:38 the ship exploded midships, caught fire, andsank. Debris was scattered everywhere, cutting an on-shore cargo line which fed fuel to the fire.Nine people were killed on board. At least, 58 people were injured, many of them on shore. Over400 boats were oiled. Millions of dollars of shoreside damage. The USCG concluded that explosivevapors had built up in the area between the deck houses, the vapor had somehow gotten into themidship house where there were plenty of possible sources of ignition, caught fire, and flashed backto the tanks. In short, a near repeat of the British Crown. The fact that the vent piping on deckwas badly corroded and holed exacerbated matters.7 Because of the location and shoreside injuriesand damage, the Sansinena was a very high profile casualty. In 1977, the USCG required IGS onall oil tankers trading to the USA down to 20,000 deadweight tons.

    Once again the regulation only applied to ships trading to the US. IMO and the ClassificationSocieties did nothing to prevent a rash of tank explosions in the 1970s as Table 4 shows. At least510 people died and 956,000 m3 of oil was spilled in these casualties. One of the worst of these wasthe E M Queeny/Corinthos collision in which the non-inerted Queeny slid into the non-inertedCorinthos at very low speed, penetrating only one tank. The Corinthos had almost completeddischarge at the time. But the fire from that tank quickly spread to the rest of the ship, killing 24.Ironically, the Corinthos was discharging at Marcus Hook.

    Even worse was the Betelgeuse. The Betelgeuse exploded at 01:00 on the night of January8th, 1979. while discharging at Bantry Bay. All aboard and at the terminal were killed, so we dontknow exactly what happened. But we do know from the Irish investigation that:

    1. The Betelgeuses hull steel was in horrible condition, especially the midships permanent ballasttanks. Both the owner, Total, and the classification society, Bureau Veritas, was aware of this,but had done nothing in part because Total intended to sell the ship.[12]

    2. Ten years after the Marpessa et al this 120,000 ton ship, owned by a major oil company,was not inerted. Total tankers did not trade to the USA, so they had no need to comply withUSCG regulations.

    Best guess is that cargo leaking into the ballast tanks, was ignited, perhaps by the structure failing.Without inerting the fire quickly spread to the cargo tanks, which blew up, killing 50.

    On the plus side, we began to see some of the benefits of inerting. The Venpet/Venoil wasa high impact collision off South Africa between two 300,000 ton tankers, one of which was loaded.Both ships were inerted. The loaded Venpet travelling at 13.5 knots hit the Venoil in ballast,also at about 13.5 knots, just forward of the bridge. The Venoil caught fire aft; which fire spilledonto the Venpet, whose forward cargo tanks were holed. Both ships were abandoned. As the shipdrifted, both fires went out. Both ships survived. Only two crew were killed, The great bulk ofthe Venpets cargo was saved, and a 300,000 ton plus spill was averted. Contrast this with theOswego Guardian/Texanita a somewhat similar collision in the same area. The non-inertedTexanita exploded killing 43, despite the fact that like the Venoil she was in ballast.

    In July, 1980, about six months after the Betelgeuse the Energy Concentration brokein two while she was discharging at Rotterdam. But unlike the Beteleguese, there was no fire, nofatalities, and nil pollution. The Energy Concentration was inerted.

    6 The regulation applied to pure tankers over 100,000 dwt and combination carriers over 50,000 dwt whose keel-laying date was after 1974-12-31.

    7 The Sansinena was Liberian flag, classed by Lloyds Register.


  • Table 4: 1970s Casualties where inerting probably would have helpedBased on CTX CDB Version 4.4 as of 2010-03-11T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    19701023 pacific glory M 13 6000 Pac. Glory/Allegro in Channel, one said overtaking, one crossing19710111 texaco caribbean P 8 638 fog, coll w rogue vessel E.C. exploded, sank, others hit wreck19710217 ferncastle M 7 0 tank explosion in ballast, sank, no real cause info19711115 elcano M 4 0 another non-inerted tank cleaning explosion, 4 killed19720128 golden drake M 2 32000 explosion forward while loaded, sank, Bulkpetrol Class, cause?19720201 v a fogg P 39 0 T2 explosion tank cleaning off Texas, part-loaded, no IG19720420 silver castle M 16 0 coll w S.A. Pioneer in dense fog, no cause, lightered, sunk19720506 esso chittagong M 0 0 slow speed dance of death at Bangkok pilot station19720511 tien chee M 8 5600 collision w Royston Grange Rio Plata due to bank effect, 82 dead19720514 golden jay M 3 0 tank explosion, sank in ballast, probably hotwork in bad tank19720821 texanita M 43 0 coll w Texanita off S. Africa, Texanita explodes, sank,cause?19720826 princess irene P 6 0 lightning strike discharging crude Donges, not inerted, 6 killed19721101 san nicolas P 3 0 tank exp leaving Brindisi after discharging naptha, no inerting19721216 bello M 0 18000 explosion W Med in No 3 tank, prob sbt, towed in, scrapped19721219 sea star M 12 141100 collision Gulf Oman, one Port to Port, one stbd to stbd19730602 esso brussels M 13 5000 rammed NY harbor by Sea Witch whose steering gear failed19730621 mobil pegasus Y 0 0 non-inerted 1C exploded during ballast exchange, abt 5 hurt19730625 philippine leader M 6 0 tank exp in ballast off S Africa, 6 killed, no cause info19731105 golar patricia Y 1 5880 explosion off Canaries in tank being cleaned, sank19740118 keytrader P 2 2790 dance of death with Baune in lower Mississippi River19740222 nai giovanna 8 3490 OBO, tank explosion N Pacific, tank clning?, Berge Istra, Vanga?19740409 elias P 13 3500 tank explosion, discharging Delaware R no inert, rust, no cause19740809 metula M 0 62000 grounded Str of Magellan, pilot error, no place for VLCC19741010 tekton M 0 600 hit by loaded OBO Queen in fog off Pt Elizabeth, sank, cause?19750129 jakob maersk M 6 98800 grounding?, massive ER explosion mooring Leixos, cause???19750131 corinthos M 23 42200 hit by E M Queeny, Marcus Hook, no IG, no twin screw, pilot error19750131 e m queeny M 1 0 hit by E M Queeny, Marcus Hook, no IG, no twin screw, pilot error19750417 tosa maru M 0 2000 hit by loaded Cactus Queen off Spore, sank, no cause info.19750417 cactus queen M 0 0 hit by loaded Cactus Queen off Spore, sank, no cause info.19750815 globtik sun M 6 1110 hit platform off Galveston, bad charts, bad plotting19751028 kriti sun P 0 3490 fire Singapore SBM, lightning, just after discharge, inerted?19751229 berge istra P 30 5000 explosion in OBO double bottom Celebes Sea, 30 of 32 killed19760512 urquiola M 1 111700 grnd on uncharted ledge, ordered out, grnd again, fire, sank19761217 sansinena P 9 47600 explosion on non-inerted deck at LA, ballasting, 9 killed19770223 hawaiian patriot M 1 115000 big hull crack off Hawaii, fire, broke in two, sank19770320 claude conway P 12 0 tank explosion (cleaning?), broke in 2 off Cape Fear19770627 gunny P 2 0 tank explosion discharging Sete, hull failed, inerted?19780223 cassiopeia P 5 0 tank fire in ballast, prob cleaning, prob no IG, but need data19781108 feoso sun M 30 1000 Massive explosion inspecting damage after discharge, Manila Bay19790101 master michael 31 6500 fire, sank, 31 of 35 lost, nil info19790108 betelgeuse P 50 47000 bad rust in uncoated ballast tank, no inert, explosion, 50 dead19790227 saint chris M 1 0 OBO tank cleaning explosion, East Coast USA, 1 dead, nil info19790228 aviles M 12 12000 broke in two Arabian Gulf, fire, sank,19790419 seatiger P 2 0 explosion dsching Texas, sank, lightning, IG off for cargo survey19790626 vera berlingieri M 1 6000 collision in fog w Emmanuel Delmas, fire, cause?, 29 dead, sank19790816 ioannis angelicous 4 37600 explosion just after loading Malongo SBM, sank, cargo survey?19790901 chevron hawaii P 3 32000 explosion Deer Park, lightning combined with no/poor inerting19791020 talavera M 1 0 Tank explosion in ballast, CTL, non-inerted tank cleaning?19791027 gunvor maersk M 0 20000 struck submerged object Amazon, fire, sank, need more info19791029 berge vanga P 40 5000 Repeat of Berge Istra, OBO dbl bottom explosion, 40 dead19791101 burmah agate M 32 40500 collision w Mimosa inbound Galveston Bay, 30+ dead, no detect?19791213 energy determinati P 1 0 explosion in slop tank Hormuz, inerting not in use, sank


  • 5 The IMO Finally Acts

    Non-inerted tank explosions continued into the early 1980s. In early 1980, we had three VLCCtank cleaning explosions within a few weeks of each other: Maria Alejandra, Mycene, and theAlbahaa B.

    Table 5: 1980s Non-War Casualties where inerting probably would have helpedBased on CTX CDB Version 4.4 as of 2010-03-11T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    19800223 irenes serenade 2 120000 fire lowering anchor Pylos, prob cgo in FP tank19800311 maria alejandra P 36 4660 massive explosions off Mauritania, tank cleaning?, 36 killed19800403 albahaa b P 6 4660 tank fire off Tanzania while tank cleaning, very likely bad IG19800403 mycene P 1 4660 tank fire off Sierra Leone, prob bad IG while tank cleaning?19800722 energy concentrati P 0 0 broke back dschging Rotterdam, tired CO screw up, IG worked19800821 texaco north dakot M 0 2860 hit new, unlighted jacket in GOM, charts not up-to-date19801231 blossom M 1 0 explosion off Sardinia, broke in two midships cause?19810329 cavo cambanos M 6 24300 fire in generator room Tarragona, fire, later scuttled, cause?19810520 anna xyla M 0 0 pump room, engine room fire, Jebel Dhanna, beached, no cause info19810531 monticello victory P 0 0 corroded bilge line/hose to cgo tank left open, ER exp. Pt Arthur19810712 hakuyoh maru P 7 3490 struck by lightning end discharge Genoa, probably no IGS?19811210 riva i M 2 0 Accom? fire discharging Eilat, scrapped, no other info19820306 golden dolphin P 9 3490 hot work on deck, tank explosion,19821013 unirea M 2 757 tank explosions off Bulgaria, sank, tank cleaning? inerted?19830107 assimi M 0 60200 ER fire, Gulf of Oman, cause uncertain, spreads to tanks, sinks19831126 pnoc basilan M 6 19800 waves forced open a tank cover, ? fire off Luzon, sank19840226 american eagle M 8 582 exp cleaning gasoline tank GOM, static from plastic sleeve19850408 fuji M 2 582 tank cleaning exp, prob no IG, bow sank, stern CTL19850526 petragen one M 14 5880 explosion San Roque refinery Algeciras while discharging naptha19850526 camponavia M 12 0 explosion San Roque refinery Algeciras while discharging naptha19850914 sinoda P 1 582 explosion while gas-freeing off Japan19860303 galini Y 5 0 headed for repairs, tnk clning, two explosions, 5 killed, CTL19860507 alexandros f P 0 0 OBO loaded iron ore, No 3 hold explodes, sank, crew rescued19860604 southern cross M 3 9320 tank explosion while loading gasoil Skikda, cause? inerted?19861001 angel M 0 0 exoceted by Iraqis in ballast, hit fwd, fire, repaired19870623 vitoria M 6 20 collision w Vitoria in Seine damage to helm19880422 athenian venture M 29 40000 fire off Newfoundland, broke in 2, hull failure?19880831 fiona 1 0 exp,static chg from stm leak, no IG, set off by tape, Long Is.19890314 maasgusar P 23 0 Explosion in ER, spread to methanol cargo, broke in two sank

    At this point, the underwriters at Lloyds finally stepped in and imposed a surcharge of 0.1%per year on VLCCs not fitted with IGS.[7, page 158] Assuming a $20,000,000 dollar hull value, the$20,000 per year saving would pay for the system in two or three years. By 1983, almost all bigtankers were fitted with IGS.

    This was fortunate for in late 1982, the Iran/Iraq War started and tankers were caught in themiddle. Over the next six years, there were at least 346 attacks on defenseless tankers in the PersianGulf. At least 198 tanker men were killed. But the toll would have been far worse without inerting.Only 4 tankers over 5,000 dwt were actually sunk, and in 3 of these sinkings, theres a good chancethe ship was not inerted. In 286 of the 346 attacks, the tanker survived to trade again.8 Heres aparticularly dramatic example of the value of inerting from Newton.[11, page 118]

    A typical attack on a tanker is recounted by Captain Bruce Ewen, master at the time of the412,000 dwt World Petrobras which was bombed by Iraqi jets on 22 December 1987. At thetime the tanker was providing floating storage off Irans Larak Island in the northern part of theStrait of Hormuz. Two Russian made 500 lb bombs with parachute drogues attached droppedonto the maindeck during the attack by Mirage jets, which also hit two other tankers off theisland.

    World Petrobras was at the time transferring oil from one tanker, Free Enterprise, into another,British Respect. When the bombs struck, Ewen recalls, the rubber hoses attaching us to theBritish Respect were set afire and a large amount of shrapnel from our deck fittings blew throughthe side of the British Respect. Since we were both inerted and had our inert gas plants running,an explosion was avoided. [Emphasis mine.] However, we needed to get British Respect awayfrom us so we could get firefighting tugs alongside.

    We cut her aft ropes and her master went ahead on the engines and ran the forward ropes offthe reels. When she parted the hoses, a large amount of oil was dumped into the water which

    8 These statistics were extracted from Navias and Hooton.[10] However, I am responsible for the numbers.


  • caused a large fire and set the rubber fenders ablaze. Although this rendered our lifeboat andthe liferaft on the port side beyond use, the current was fairly quick so the danger passed in afairly short time.

    The World Petrobras resumed operations 42 hours later.There are major differences between the World War II and Gulf War attacks, mainly the size

    of the ships. But still the differences are striking. In the one case, we know of 264 casualties whichkilled about 5000 crew; in the other, we know of 346 casualties which killed about 200.9

    During this period, the International Maritime Organization (IMO) finally made an appearance.IMO rules required IGS on all new oil tankers above 20,000 deadweight delivered after May 1982.By May, 1985, these rules were extended to existing tankers. Chemical tankers were exempted.as were product carriers under 40,000 dwt which were not fitted with high capacity tank washingmachines, Chemical tankers were exempted on the grounds that the inert gas would contaminatechemical cargos. No rationale was offered for the the 20,000 dwt lower limit.

    For the most part, these IMO rules merely ratified what had already happened. In any event,they were about 50 years too late. However, the IMO Regs did include one important newish safetyfactor, and that was a requirement for a closed ullage system. This simple, cheap rule allowed cargosurveys to take place without opening up the tank and destroying the inert atmosphere. The CTXdatabase contains 8 casualties involving 24 deaths and 44,000 m3 of spilllage, where we believeopening up the tanks for survey may have been a causal factor. IGS combined with closed ullageand modern tank cleaning machines would have prevented the first Bidwell explosion.

    Table 6: Casualties where opening tanks for cargo survey was probably a factorDate Ship SVY Dead Vol(m3) Synopsis

    19720826 princess irene M 6 0 lightning strike discharging crude Donges, not inerted, 6 killed19751028 kriti sun M 0 3490 fire Singapore SBM, lightning, just after discharge, inerted?19790419 seatiger Y 2 0 explosion dsching Texas, sank, lightning, IG off for cargo survey19790816 ioannis angelicous P 4 37600 explosion just after loading Malongo SBM, sank, cargo survey?19810712 hakuyoh maru P 7 3490 struck by lightning end discharge Genoa, probably no IGS?19880831 fiona Y 1 0 exp,static chg from stm leak, no IG, set off by tape, Long Is.20010117 emilia theresa Y 0 0 bad survey benzene fire after loading Santa Clara Brazil20030619 efxinos M 4 0 explosion at end of lightering off Fujairah, cause ?

    9 Moore lists 145 attacks on American flag tankers in World War II. 107 of these were sinkings. Of the 38 thatwere not, 5 were Sun Oil tankers.[9]


  • 6 The 1990s: most of the fleet finally inerted

    The value of inerting really shows up in our rather short list of tanker explosions in the Nineties.

    Table 7: 1990s Non-War Casualties where inerting probably would have helpedBased on CTX CDB Version 4.4 as of 2010-03-11T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    19910411 haven M 5 16470 tank fire anch off Genoa, prob leak into SBT, 2C19920418 world hitachi zose P 1 900 collision with bulk carrier off Morocco, holed, fire 1S, cause?19920420 seastar P 2 0 tank vented during repair work forward, explosion in 3C,IG?,19920919 nagasaki spirit P 26 14100 piracy, NUC collision w Ocean Blessing N end of Malacca, 51 dead19930925 altair P 3 0 explosion during manual tank cleaning, tank not inerted19940313 nassia P 18 23500 BC Shipbroker black out, no rudder, coll Bosporus, massive exp19981029 champion trader M 1 12 hotwork abv bfo tank, 1 dead, small BFO spill, 3000 bbl palm oil

    The Haven probably involved cargo leaking into into a non-inerted pemanent ballast tank,combined with poor inerting.

    The Nagasaki Spirit was actually a victory for inerting. The container ship Ocean Blessingwas captured by pirates at the west end of the Malacca Straits. The crew either locked themselvesin or were locked in. The ship proceeded at high speed eastbound, apparently out of control. At21 knots, she T-boned the 95,987 dwt loaded tanker, Nagasaki Spirit, on the port side aft. Thebreached tanks on the Nagaski Spirit caught fire and the fire spread to the non-inerted OceanBlessing which was engulfed in flames. The entire 25 man crew of the Blessing and probably someof the pirates were fried. These deaths are not included in Table 7 since they could not have beenavoided by controlling the fire on the tanker. But on the inerted Nagasaki Spirit, the fire wasconfined to the port side aft. It appears the crew successfully abandoned ship in the starboardlifeboat; but were then murdered for their wallets and watches. The Nagasaki Spirit survived andwas repaired.

    The Altair is a bit of an inigma. She blew up tank-cleaning. Her last cargo was naptha. Weknow she was not inerted at the time; but we are not sure whether she was fitted with an IG system.With a deadweight of 20,848 tons, she was slightly above the arbitrary 20,000 dwt limit; but mayhave been exempted from inerting if her tank washing machines were small enough, What we doknow is that she was detained two months earlier by the Canadian Coast Guard in Victoria. TheCanadians found the ship in very poor condition. In any event, inerting is useless unless it is used.

    We dont know why the death toll in the Nassia/BC Shipbroker collision was so high. Bothships were Cypriot flag. The flag state investigation, if there was one, has not been made public.We do know that the BC Shipbroker had a black out in the Bosporus, lost steering, and turnedinto the Nassia rupturing at least 2 forward port cargo tanks. The fire immediately spread to theShipbroker, killing 24 of her crew.10 The fire also apparently spread aft on the Nassia, killing 18 ofher crew, but most of the cargo was unharmed. Not much of a victory for inerting.

    But the overall performance in the Nineties when compared to the Eighties, when the fleet waspartly inerted, and especially to the Seventies, when the fleet was largely uninerted, is a very strongtestament to the efficacy of inerting.

    7 2000 to the present

    So far in the 21st century almost all tanker explosions have involved ships less than 20,000 dwt orparcel tankers which were exempted because they can carry chemical other than petroleum.

    The exceptions are the Limburg, Heng San the Efxinos and the NorthseaThe Limburg is actually a victory for inerting. This 299,364 dwt VLCC was bombed by

    Jihadists off Yemen while partially loaded. Thanks to her inerting, the fire was contained and theship survived.

    The Heng San, a 24 year old small VLCC, had extra people on board for tank cleaning, andmaybe tank repairing. The ship was ventilating 4C, which apparently still had a little cargo in it

    10 These deaths are not included in Table 7.


  • Table 8: 2000-2010 Casualties where inerting probably would have helpedBased on CTX CDB Version 4.4 as of 2010-03-11T08:25:19

    Date Ship IGS Dead Vol(m3) Synopsis

    20010115 p harmony 9 500 unsafe charter forced unsafe purging, tnk explosion, no IG, sank20010611 heng san 7 0 tank explosion while purging Arabian Sea, 7 killed, sinks?20021006 limburg M 1 0 attacked while approaching Yemen SBM to top off20030613 chassiron 1 0 tank explosion during tank cleaning, not inerted, one killed20030619 efxinos 4 0 explosion at end of lightering off Fujairah, cause ?20040101 panam serena P 0 0 dschging benzene Porto Torres, no inert, fire, 2 killed20040228 bow mariner P 21 12600 non-inert tank cleaning, explosion off Virginia, sank20040404 ncc mekka P 2 0 washing non-inerted tank off Brazil, fire, two killed20041115 vicuna P 6 5000 no inert?, tank explosion discharging methanol Paranagua, sank20041215 sunny jewel 3 0 no IGS, poor tank purging, explosion while tank clning, 3 killed20070529 northsea M 7 0 31 yr old SH, claimed geg-oil tanker, after lightering Nigeria

    with steam driven fans and long ducts, when the tank exploded. The Singapore flag state reportedthat the inert gas system was not operating, but does not say why. In short, the Heng San provedthe obvious; IGS cant do any good if its not used.

    The 57,372 dwt Efxinos exploded after discharging a cargo into a bigger ship off Fujairah. Theflag state (Malta) report is secret, so we dont know what happened. Despite the closed ullagesystem requirement, she may have opened her tanks for post-discharge cargo survey.

    The Northsea is a wild one. This 32,290 dwt tanker exploded killing seven, just after reverselightering off Nigeria, allegedly struck by lightning. (There is no economic reason for reverse lighter-ing in this area.) The ship was a 31 year old, pre-Marpol single hull trading oil way past the singlehull deadline. She is listed as a vegetable oil tanker, which she clearly wasnt. She was delistedby Lloyds in 2005; there is no record of any replacement Class. At the same time she shifted toCambodian flag. Given all the other irregularities, we have no idea if the Northsea was inerted.Even if the lightning story is true, lightning should have nil effect on a properly inerted tanker. Ican only conclude that 73 years after the Bidwell, this ship was effectively un-inerted.

    In any event, with only four big tanker explosions, regulatory attention turned to the arbitary20,000 dwt limit and the exemption of chemical tankers. The main motivation was the Chassiron,Panam Serena, and Bow Mariner casualties.

    The 9,995 dwt chemical tanker Chassiron was exempted from inerting. She was tank cleaningafter discharging a cargo of heating oil (1 P&S), gas oil (2,3,4,5 P&S), and 98 octane gasoline(6 P&S). Just after tank cleaning started in the 6s, a whistling sound was heard, followed byan explosion in 6S, then 6P exploded, and then 5P. One crewman was killed. The investigationconcluded that oxygen had entered all the tanks during discharge and opening the tanks for cargosurvey, resulting in a highly explosive atmosphere in at least the 6 across tanks. The source ofignition could not be determined with certainty but most likely was the in-tank cargo pump. Theinvestigators pointed out that the atmosphere in 6S would have required only a few microjoules ofenergy to ignite.

    Six months later, the non-inerted 7450 dwt Sun Venus exploded in the Inland Sea, killing two.She was cleaning tanks after discharging benzene and alcohol.

    A year later, the non-inerted 6880 dwt Sunny Jewel suffered a somewhat similar tank cleaningexplosion off Japan, killing 3 crew members and sinking the ship.

    The 10,047 dwt Panam Serena was discharging benzene at Porto Torres. At 11:55, there wasthe first of a series of four explosions killing two. The time between explosions was approximatelyone minute. The investigators were unable to determine the ignition source but it is clear that anexplosive atmosphere prvailed in the tanks and possibly on deck. Like the Chassiron, the PanamSerena was fitted with deep well pumps inside each cargo tanks. These could have been the sourceof ignition. The Panam Serena was exempt from inerting.

    The Bow Mariner was a 39,891 dwt chemical tanker which was carrying ethanol when sheexploded and sank off Virginia, killing 21. She had part-discharged MTBE in New York, and the22 empty tanks had been opened on the Captains orders for cleaning. Despite the fact that shewas exempt from inerting because she was not carrying oil, she was fitted with an IGS. But itwas not being used, which was both legal and standard operating procedure on this ship. The


  • USCG determined that the atmosphere in the opened tanks or on deck was explosive; but could notpinpoint the source of ignition. It might have been simply a spark from the stack, or more likelycaused by the crew lowering air driven pumps into the gas dangerous space, using Scuba gear! Therewere two main explosions less than two minutes apart, each of which was actually a rapid series ofexplosions, representing individual tank blowing up. The investigating USCG officer recommendedthat inerting be made mandatory for chemical carriers. But the Commandant refused to do thisclaiming the IG might contaminate chemical cargos.

    Two months later the 37,272 dwt chemical tanker NCC Mekka had a somehwat similar casualtyoff Brazil. The ship had just part-discharged paraffinic solvent in Santos and was proceeding to theUSA with the remaining alcohol cargo. She had started tank cleaning the newly emptied tanks.The ship was fitted with IGS, but like the Bow Mariner it was not being used since the cargoes werenot deemed to be oil. This despite the fact that the petroleum based solvent had a flash point of-40C! It is believed that friction in the tank cleaning machine was the source of ignition. The crewdid a great job of responding to the fire, saving the ship. Two crew were killed.

    In 2009, at IMO Norway and other argued that there was no rational for either the deadweightlimit nor the exemption of chemical tankers. Not only do chemical tankers often carry volatilepetroleum products (see Chassiron and Panam Serena); but some of the other chemicals theycarry are even more flammable than oil (see NCC Mekka).

    The argument that inert gas will contaminate chemical cargoes is easily countered. CO2 is farless reactive than O2, so the concern is water, sulfur and particulate matter. All three can be reducedto ambient levels and below by slight modifications to the IG system. Exxon and others did it withsub-cooling.[1, page 196] MSTS did it with electrostatic precipitators.[1, page 210] Hellespont didit with double scrubbing. See Appendix A. These modifications add 20% or less to the cost of aninert gas system. In the 1950s, Sun Oil demonstrated to the US Navy that JP-4 jet fuel which hasvery strict specs could be carried in inerted tanks without these modifications.[2, page 240] Thecontamination argument is simply a stalling device.

    Norway and its allies argued that all cargoes with a low flash point represented the same hazardand should be treated in the same manner. However, Norways argument was undermined by twoFormal Safety Assessments (FSAs) which concluded the cost of implementing inerting to theseships were more than the dollar benefits of the lives and ships saved. One of these FSAs was doneby Norway.

    In early 2009, the IMO Sub-committee on Fire Protection was deadlocked on the issue, andextended the target data for new requirements to 2011. Seventy-eight years after the Bidwell,tankers below 20,000 dwt or carrying chemicals are not required to be inerted. The Bidwell hada deadweight of 10,950 tons.

    8 May They Rest in Peace

    Table 9 summarizes the CTX CDB non-War tanker casualties in which IGS would probably havehelped by period. We know this table is terribly incomplete prior to 1960 and probably badlyincomplete in the Sixties. It should be reasonably complete for major fires from 1970 on.

    Looking at Table 9, it is hard to avoid the conclusion that, if the industry had responded asdecisively as Sun Oil did in 1932, at least several hundred, perhaps as many as 500 fiery deaths,would have been avoided. And that does not include the potential for drastically decreasing the5000 or so allied tankermen killed in World War II. To put it another way, without inerting wewould probably be killing 250 more tanker crew per decade.

    The obvious question is: why was the industry so slow to adopt such an obvious, effective,safety measure which probably pays for itself in reduction of tank corrosion? Why 68 years afterthe Bidwell are we arguing whether tankers below 20,000 dwt should be inerted?

    The equally obvious answer is that tanker owners with a few notable exceptions could not bebothered. A tanker owner rarely suffers any loss when one of his ships blows up. His P&I insurancepays off the dead crews family, in most past cases a few thousand dollars per head. His hull


  • Table 9: Summary of non-War casualties where inerting probably would have helped

    Period Killed Spilled(m3)

    Pre-WW 11 88 ???1945 thru 1960 119 87,0001961 thru 1969 218 185,4261970 thru 1979 513 956,0981980 thru 1989 278 305,8431990 thru 1999 105 54,9822000 thru 2010 63 18,100

    insurance covers the loss of the ship. In many cases, the insured value is more than the marketvalue and the owner comes out ahead.

    If the owners dont care, the tanker regulatory process which is based on the Flag State and theClassification Society, wont care.11 Both Flag States and Classification Societies must compete forowners. They are in no position to impose requirements that their customers deem unnecessary.

    The International Maritime Organization is basically a collective of flag states. IMO voting andbill paying is based on the size of each members fleet. The flag states pay IMOs bills, so the bigflags of convenience are the real power at the IMO. IMOs job is to impose as little real regulationon ship owners as public opinion will allow.

    Recently, the owner/flag state/class trinity has developed a new weapon to forestall meaningfulregulation. Its called Formal Safety Assessment. All proposed IMO regulation is supposed to beevaluated by Formal Safety Assessment (FSA). FSA values a life saved at $3,000,000. Lets supposethe worlds tanker fleet was currently not inerted. We propose to impose inerting on all the 5300tankers over 10,000 dwt tons world-wide. To do this will cost on average roughly $400,000 per shipor a bit over 2.1 billion dollars for the fleet as a whole. Assume a ship life of 20 years. Accordingto FSA, to justify this expenditure, we need to save 700 lives over the next two decades. But Table9 indicates we would have difficulty arguing for a saving of more than 500 lives over that period.12

    FSA claims that the most important tanker safety measure of all time is not cost-effective, andshould not have been imposed.

    With reasoning like this, perhaps the strange regulatory history of tank inerting is not so strangeafterall.

    11 The Classification Societies have played absolutely no role in this tale, other than to approve badly corrodedsteel.

    12 Ive conveniently forgotten about the reduction in corrosion and the value in war, but a skilled FSA practitionerwould have little difficulty shooting down such speculative claims. See Formal Safety Assessment in Wonderland fora more detailed critique of FSA.


  • Appendix A. Ballast Tank Inerting

    In the 1990s at least two tanker owners, Tankship Transport and Hellespont Shipping, begainextending tank inerting to segregated ballast tanks under my direction. The motive was corrosioncontrol. Our VLCCs and ULCCs were built in the mid-1970. The ballast tank coatings on these15 year old ships was breaking down badly, especially in the top of the tanks. The idea was thatby inerting we could obtain the same corrosion control benefits we had seen in the uncoated cargotanks, some of which were also used for ballast.

    The problem was sulfur. Tanker fuel, which is the dregs of the refining process, generally contains3 to 4 percent sulfur. This means the stack gas coming out of the boilers contains 1500 to 2000 ppmsulfur dioxide. A good scrubber, properly maintained and operated, will cool the stack gas fromabout 200C to about 17C above the seawater temperature, condensing much of the water vapor inthe boiler exhaust. Much of the SO2 dissolves into the condensed water which is drained out ofthe bottom of the scrubber and discharged overboard. This process will remove 95 to 97% of thesulfur. The resulting gas still contains 50 to 100 ppm SO2. It is smoky, smells of sulfur, and it stillgenerates sulfur related corrosion, which can be seen in the IGS piping and in the cargo tanks nearthe inert gas inlets.

    We found that by installing a second scrubber in series with the first, we could removed 95% ofthe remaining sulfur reducing the SO2 content to less than 2 ppm at full load on the inert gas fans.At deballast volumes which are typically one-fifth cargo discharge rates, the sulphur content dropsdown to 0.2 to 0.3 ppm. You can get more than this on a winter day in New York city with aninversion. The double scrubbed stack gas is clear as a bell and has no smell. The reason why thesecond scrubber is so effective is that it reduces the stack gas temperature from about 17C aboveseawater to only 2 or 3C above seawater as shown in Figure 4. This condenses a great deal of watervapor in the stack gas. which in turn pulls out most of the remaining sulfur.

    Figure 4: Double Scrubbing Schematic







    1500 ppm S02200C

    50 ppm S02+17C

    2 ppm S02+3C




    Tanks2 x 100% Fans

    5:1 turndown

    The first such system was installed on Tankships Empress des Mer in 1993. This was sosuccessful it was quickly followed by Hellespont Shippings installing the system on seven UltraLarge Crude Carriers (ULCCs) ranging from 320,00 to 420,000 dwt. When we did this, I hadthe crew purposely scrape holes in the coating in about 240 spots in one of the ballast tanks, andcarefully measure the steel thickness at each of these spots. Fourteen months later we went backinto this tank and remeasured these spots.

    There was no wastage.13 You did not need the thickness gauge to tell you that. The 1 to 2cm diameter bare spots were like new except that they were covered with a very thin black film.

    13 The actual measured mean wastage was 0.009 millimeters, but this was clearly measurement noise. The standarddeviation of the measurements was 0.113 millimeters, ten times that of the mean.


  • There was no scale and no evidence of the 15 year old coating being lifted at the edges of the barespots. The tank was well-anoded so there was plenty of white calcareous deposits in the tank fromthe ballast legs.14 But what was really interesting was there was no red or brown anywhere in thetank. At the top of the tank, where the coating had broken down earlier we had had some leafingcorrosion, layers of red-brown rust that typically forms in corners. But now the leaves had turnedblack. I broke off one of the leaves and sent it to a lab. It was almost all magnetite, Fe3O4.

    Corrosion proceeds from iron to magnetite, Fe3O4, to ferric oxide Fe2O3, better known as rust.The inerting had reversed this process. We had created reducing conditions in the tank.There will be no rust in such a tank.

    In order to obtain these results, it is essential that no oxygen be allowed into the tanks. Allinerted tanks must be fitted with P/V valves. These valves release if the pressure or the vacuum inthe tank becomes too large to protect the structure. During the day, the tank contents heat up andexpand. If the increase in pressure in the tank exceeds the release point of the P/V valves, some ofthe tank contents will be released into the atmosphere. In the night, when the tank cools outsideair will be sucked into replace the released gas introducing oxygen into the tank. This is called tankbreathing. Need BP citation

    Tank breathing can be controlled by:1. Painting the decks white. With white decks, the P/V valve release points will almost never

    be reached.2. Continuously monitoring the pressure in the IGS system and turning on the blowers as soon

    as the pressure in the tank gets too close to ambient. Our rule was: if the pressure droppedbelow 0.05 bar gage, then the inert gas must be topped off at the next 0200 local. If thepressure dropped below 0.02 bar gage, the the IG is topped off immediately. For motor ships,the will require an independent IG generator.

    3. Carefully maintaining the P/V valves, which have a tendency to leak. Leakage will exposeitself in the pressure readings, and must be addressed promptly. A prerequisite here is stainlesssteel seat rings.

    It is also important that the tank vents be located at the highest point of the tank, taking accountof trim, to prevent build up of high O2 pockets in the top of the tank.

    In 2001/2002 Hellespont built four double hull ULCCs which followed all these rules. Theunderdeck portion of the cargo tanks was uncoated. When the first of these ships was delivered,some 200 spots on the deck were marked and the steel thickness carefully measured. Two yearslater, just before the ships were sold, we could measure no change in thickness. Nor was there anyvisual sign of wastage in the underdeck steel. Bare steel coupons placed in the top of the ballasttanks, also showed no sign of wastage.

    An intriguing variant or addition to double scrubbing is dehumification of the flue gas. Theeasiest way to do this is by cooling. In the early 60s Exxon (then Esso) fitted two ships withIG systems which had a cooling section in the top of the scubber.[1, page 196] This section was asimple fin tube coil fed by chilled (5C) water from the ships air conditioning system. With thismodification, Exxon was able to reduce the outlet gas temperature from 27C to 10C, roughly thesame reduction Hellespont achieved with double scrubbing. Exxon found that this reduction intemperature reduced the water vapor content in the inert gas by a factor of four.

    It is an open question whether double scrubbing or cooling is the best way to go to improvingthe quality of the inert gas. Both will work. My own view is that they should be combined whichwill result in a truly dry, clean inert gas being injected into the tanks, a gas that not even thepickiest chemical tanker operator can object to.

    14 Inert gas will not do anything for sea-water corrosion of the steel that is submerged during the ballast legs. Thismust be handled by anodes. Givn proper anoding, more electrons are pushed into the steel than are released by thecorrosion process. This can be measured by a voltmeter, and also shows up in white calcareous deposits wherever thecoating has broken down.[4, page 306-307]


  • References

    [1] Stanford A. What inert-gas systems mean to a marine operator. SNAME Transactions, 71:203219, 1963.

    [2] C. Boyle. Corrosion control with inert gas. 6th World Petroleum Congress, pages 235242,1963. June 19-26, Frankfurt, paper 10621.

    [3] Sunoco Marine Department. History of the sun oil company marine department. Technicalreport, Sun Oil Company, 2001. Title and year a guess.

    [4] Jack Devanney. The Tankship Tromedy. Center for Tankship Excellence, 2006.

    [5] P Heywood. The bp inert gas system for oil tankers. Technical report, BSRA, 1969. B.S.R.A/Report NS. 268.

    [6] C. Hocking. Dictionary of Disasters at Sea, 1824-1962. The London Stamp Exchange, 1969.

    [7] G. Hodgson. Lloyds of London, A Reputation at Risk. Penguin Books, 1986.

    [8] B. Makin. Static electrification in supertankers. Physics in Technology, 2:109116, 1975.

    [9] A. Moore. A Careless Word, a Needless Sinking. American Merchant Marine Museum, 1983.

    [10] M. X. Navias and E. R. Hooton. Tanker Wars: The Assault on Merchant Shipping during theIran-Iraq Crisis, 1980-1988. Tauris, 1996.

    [11] John Newton. A Century of Tankers. INTERTANKO, 2002.

    [12] Tribunal of Inquiry. Disaster at Whiddy Island, Bantry, County Cork. Irish Stationery Office,1979. Prl. 8911.


    The Origins of Tankship InertingNo Reaction for 30 YearsVLCC's and InertingThe USCG Finally ActsThe IMO Finally ActsThe 1990's: most of the fleet finally inerted2000 to the presentMay They Rest in Peace


View more >